ARTIFICIAL NEURAL NETWORK BASED ASSESSMENT OF GROUNDWATER QUALITY IN DELHI REGION, INDIA

Abstract

Groundwater is a crucial natural resource that supports both ecosystems and human activity. For the purpose of providing clean drinking water sources and environmental protection, it is essential to evaluate the quality of groundwater. In order to evaluate groundwater quality accurately and effectively, this study suggests a framework called the GQI that uses ANN. By capturing the intricate linkages and nonlinear patterns seen in the dataset, the ANN architecture makes it possible to forecast the quality of groundwater depending on the input parameters. The model goes through a thorough training procedure where the biases and weights are optimised to reduce prediction mistakes. A number of benefits come from using ANN in the GQI technique, including flexibility, adaptability, and the capacity to handle sizable and complicated datasets. The suggested paradigm offers a useful tool for policymakers, managers of water resources, and decision-makers to analyse and manage the quality of groundwater, enabling prompt responses to safeguard both human health and the environment. Once trained, the ANN model may be used to aggregate the normalised values of the water quality parameters and determine the GQI for every given groundwater sample. The GQI offers a thorough and succinct depiction of groundwater quality, making it simple to compare and comprehend the findings. The study has been carried out with the objective to determine the WQI for duration of 8 years i.e., 2014 to 2022 and model and predict the WQI by ANN approach. The obtained results using the ANN model in the training phase are observed to be substantial, and optimistic with extremely high association having an R-square value of 99.99% - 100% thus exhibiting the suggested program very high efficiency.